At the end of last year Meta announced an update to their machine learning capabilities. The update was named Andromeda, and it applies specifically to Advantage Plus campaigns.
Before we get started I want to thank this week's sponsor: Segments Analytics.
Built by data scientists, Segments Analytics analyzes your Shopify store to deliver actionable insights and AI-powered segmentation, allowing you to identify profitable customer segments and retain customers effectively.
Meta releases code updates periodically, so there hasn’t been a ton of industry discussion about this update. But there has been some speculation that this update indicates a “new normal” for Meta ads media buying strategies.
I’m just going to spit it out: some folks are speculating that Meta will phase out ABO/CBO and ASC campaigns will be the only option.
No one knows for sure when or if this will happen (I guess except for Zuck). In this issue I’m going to share the context you need to understand this update and my recommendations for navigating it.
Context: Campaign Types
First, some context setting to ensure you understand the rest of this issue. If you’re deep in the weeds of media buying, you can skip this part.
Meta campaigns can roughly be divided into “old algo” and “new algo”.
Old algo campaigns are CBO (campaign budget optimization) and ABO (adset budget optimization). These are the campaign types that have been around since roughly 2016.
In CBO campaigns, the budget is allocated at the campaign level. Meta distributes budget across ad sets based on the performance of individual ads. In ABO campaigns, budget is allocated at the ad set level and distributed across ads within the ad set.
You generally use ABO campaigns when you want to force spend into an ad set, or a specific ad. There are some media buyers who believe that it’s impossible to validate an ad until it receives a certain dollar value of spend.
Both CBO and ABO campaigns allow you, the advertiser, to define specific audiences for targeting and exclusion.
ASC campaigns are slightly different. There is only one ad set per campaign. You, the advertiser, are not able to define targeting and exclusion audiences.
Instead, you set a targeting cap for existing/engaged customers (more on that later). If you want to run a pure prospecting campaign, you set the existing/engaged customer cap to zero.
Both ABO and CBO campaigns match ads with relevant users via the “old algo”. ASC matches ads with relevant users via the “new algo” aka Andromeda.
***
If you're focusing all your energy on acquiring customers, but no energy on retaining customers, you are leaving money on the table.
Brands with stronger repeat rates and LTV can scale paid acquisition harder, with more confidence. LTV = cash flow. But many fashion brands struggle with retention because big assortments translate into thousands of potential purchase journeys.
There is a solution: Segments Analytics uses powerful analysis and predictive analytics to help you cut through the noise and identify your highest value customers.
Segments Analytics makes it easy to identify high value segments, find the products that are most likely to convert them, and push personalized content to Meta, Klaviyo and all your other marketing channels.
Brands with big, broad product assortments–brands like ILIA Beauty and Jambys–use Segments Analytics to maximize the ROI of their retention marketing, lift repeat rates and increase customer LTV.
Click here to try Segments Analytics free for 14 days.
***
I’m not a data scientist. I don’t know the technical differences between old and new. I would venture a guess that only folks who have worked on the Meta engineering team actually know for sure.
If you are a data scientist, or really interested in AI, this X thread has a good breakdown of the technical aspects of the update. It does not extrapolate these into media buying recs; this is pure nerd out content.
Many media buyers and brand owners complain that ASC campaigns tend to favor existing customers and highly engaged users. They go after “low hanging fruit” first and are harder to break out of a local maximum w/r/t targeting.
Making The Most Of Andromeda
Nate Lorenzen, the founder of Dysrupt agency, shared three actionable takeaways for working with the Andromeda update:
1. Increase your creative volume. Andromeda favors advertisers that give it more data points (truly diverse ad creatives) to work with. Click here for my creative diversity playbook for fashion brands.
2. Embrace broad targeting. Even if you’re running “old algo” campaigns, you’ll get more scale if you put fewer targeting conditions in place.
3. Optimize your entire path to purchase. Post-click signals (how the user behaves on your website after clicking the ad) will matter a lot more. Poor web conversion rates will be more detrimental to ad performance.
There are a few other, more basic things that I think fashion brands should focus on to optimize performance post-Andromeda:
1. Start experimenting with ASC campaigns if you haven’t already. ASC is a great vehicle for scaling your winning ad creatives. In next week’s members-only issue, I’m going to outline specific ad account structures I like. Click here to become a member.
2. Start experimenting with Advantage Plus Shopping campaigns aka feed-based dynamic catalog ads for ASC. To optimize fully, enable Meta Shops and make sure that your product feed is passing accurate info to Meta.
3. Make sure your engaged customer audience definition is optimized for prospecting. This is the audience that Meta uses to define the “engaged/existing customer exclusion” in ASC campaigns. The default setting here is sub-optimized for prospecting. This video has step by step instructions for how to do this.
I'll add one more thing: the Meta algo has been weird the past couple of weeks. If the strategies above aren't working for you, it's always worth testing out lookalikes of your best customers and interest stacks.
Rule of thumb: if an account structure/media buying approach that has been working for you for months suddenly stops working, it's worth testing different strategies.
If you don't have consistent traction yet, it's best to use the media buying strategies I outlined here and focus on creative testing, product development and site experience.
Will CBO & ABO Go Away?
No one knows for sure. There was speculation that CBO and ABO would be phased out by the end of 2024. Obviously, that didn’t happen.
Personally, the potential for CBO and ABO to go away feels more real with this update, so I would heed the recommendations above to start future-proofing my media buying strategy. You don’t have to kill winning strategies. Simply start scaling with these new tools.
One thing is crystal clear: outdated, hyper-segmented media buying strategies are going to be less effective than ever. Here’s a post on X that illustrates what I mean when I say “outdated media buying strategies”.
If this is how your account looks now, you have to bring your strategy into the future. By the way, if you'd like me to audit your Meta ads account and provide in-depth feedback, you can book an audit here.
Next week in the members-only issue I am going to outline specific account structures I like for fashion brands, along with the pros and cons for each. Click here to become a member.